基于PNN和改进D-S证据理论的夏季蛋鸡舍环境评价方法  

Environmental assessment method of hen house based on PNN and improved D-S evidence theory

作　　者：谢苗苗 李华龙 詹凯[3] XIE Miao-miao;LI Hua-long;ZHAN Kai(Hefei Institute of Technology,Anhui Hefei 230031,China;Institute of Intelligent Machines,Hefei Institutes of Physical Science,Chinese Academy of Sciences,Anhui Hefei 230031,China;Institute of Animal Husbandry and Veterinary Medicine,Anhui Academy of Agricultural Sciences,Anhui Hefei 230031,China)

机构地区：[1]合肥理工学院,安徽合肥230031 [2]中国科学院合肥物质科学研究院智能机械研究所,安徽合肥230031 [3]安徽省农业科学院畜牧兽医研究所,安徽合肥230031

出　　处：《齐齐哈尔大学学报(自然科学版)》2025年第1期36-42,76,共8页Journal of Qiqihar University(Natural Science Edition)

基　　金：安徽省质量工程一般项目(2022jyxm099);国家自然科学基金项目(31902205);国家现代农业产业技术体系(CARS-40)。

摘　　要：为实现多因子耦合复杂蛋鸡舍环境质量的评价,提出基于概率神经网络(PNN)和改进D-S证据理论的评价方法。先对蛋鸡舍环境因子数据进行预处理,构建多环境因子PNN评价模型,将模型求和层的结果归一化,作为改进D-S证据理论的基本概率分配(BPA)输入,根据各BPA的贴近度相似矩阵,得到证据支持度和可信度矩阵,再对证据加权修正,利用D-S理论合成规则迭代融合,获得蛋鸡舍环境质量的评价结果。利用夏季层叠式笼养蛋鸡舍实测数据对所建模型验证。试验表明,当PNN评价模型的结果冲突时,改进D-S证据理论对蛋鸡舍环境质量评价结果的支持率为0.8083,比传统D-S算法提高5.2%;当PNN评价模型的结果不冲突时,改进D-S证据理论对蛋鸡舍环境状态的支持率为0.9986,与传统D-S算法计算结果相同。In order to evaluate the environmental quality of multi-factor coupled complex laying hen house,an evaluation method based on probabilistic neural network(PNN)and improved D-S evidence theory was proposed.Firstly,the environmental factor data of layer hen house were preprocessed,and a multi-environmental factor PNN evaluation model was built.The results of the model summation layer were normalized,which was used as the basic probability allocation(BPA)input of improved D-S evidence theory.The evidence support and credibility matrix were obtained according to the proximity similarity matrix of each BPA,and then the evidence was weighted and revised.The D-S theory synthesis rule was iteratively fused to obtain the evaluation result of the environmental quality of the laying hen house.The measured data of the stacked cage layer hen house in summer were used to verify the constructed model,and the experiment showed that when the results of the PNN evaluation model conflict,the improved D-S evidence theory had a support rate of 0.8083 for the environmental quality evaluation result of layer hen house,which was 5.2%higher than the calculation result with the traditional D-S algorithm.When the result of the PNN evaluation model do not conflict,the improved D-S evidence theory had a support rate of 0.9986 for the environmental status of hen house,which was the same as the calculation result of the traditional D-S algorithm.

关 键 词：PNN 改进D-S证据理论 蛋鸡舍 环境质量 综合评价 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]